Introduction to Mixed Model QTL mapping
Organizers
The course is organized by the Department of Applied Statistics of Wageningen UR (Biometris)1, the
Generation Challenge Program-Integrated breeding platform2, and the Universidad Nacional de Córdoba,
Cátedra de Estadística y Biometría de la Facultad de Ciencias Agropecuarias3.
1
http://www.biometris.wur.nl/ 2 http://www.generationcp.org/ibp 3 http://www.agro.unc.edu.ar
Instructors
Marcos Malosetti: WUR-Biometris, Wageningen, The Netherlands.
Lucía Gutierrez: Departamento de Estadística, Universidad de la República, Uruguay.
Mónica Balzarini: Cátedra de Estadística y Biometría, Universidad Nacional de Códoba. CONICET.
Cecilia Bruno: Cátedra de Estadística y Biometría, Universidad Nacional de Códoba. CONICET.
When and where?
Tuesday 13th to Thursday 15th December, 2011. Monday 12 th and Friday 16 th extra-classes will be
offered as optional. The venue is the Universidad Nacional de Córdoba, Argentina, Agricultural College,
Postgraduate School (Av.Valparaiso s/n cc 509 Ciudad Universitaria, Córdoba, Argentina; Phone 54-3514334105 int 217 or int 219)
For whom?
Graduated students and professionals interested in a flexible QTL mapping approach, applicable in
simple situations (single trait, single environment) as well as more specialized situations as multiple
environments (QTLxE), multiple traits, and association mapping.
Language
Spanish
Overview of contents
QTL mapping is presented as a natural extension of the analysis of (field) trials within a mixed model
framework. Effectively, on a genomic grid genetic covariates are fitted that represent contrasts in
unobserved QTL genotype probabilities given marker information. The calculation of these genetic
covariates will be explained for various types of populations: inbreeders, outbreeders and association
panels. We will start by refreshing basic concepts of QTL mapping in the simplest situation, a single trait,
single environment QTL mapping. The mixed model application will be introduced in the context of single
traits field trial analysis and then extended for QTL detection in multiple environments (QTLxE analysis),
where the modelling of the variance-covariance structure for trials and traits is crucial. The mixed model
will be finally illustrated in the context of association mapping, where other sources of genetic variation,
this time caused by the heterogeneous genetic relatedness between individuals in the population should
be accounted for.Practical QTL analyses will be presented mainly using the graphical user interface of
GenStat 14th. In addition, equivalents (or approximations) to the GenStat mixed model formulations will
be presented in the package R and the part of the analysis in Info-Gen. It is recommended that
attendants have some familiarity with mixed models and quantitative genetics. However, on
Monday 12th an optional class day will be offered to introduce mixed model theory as related
with quantitative genetics.
Learning Objectives
By the end of the course you should be able to a) Perform a QTL analysis for a wide array of single
populations, for single and multiple environments, and single and multiple traits, b) use various
inference procedures for assessing QTL evidence, c) report QTL locations and effects.
Instruction methods
Theory will be presented in the form of lectures. Supervised practicals will allow attendants to become
familiar with the details of the actions required to perform a QTL analysis in GenStat using Windows’
dialogues, in R using source code and Info-Gen using a menu interface. These practicals will also serve
to learn how to interpret QTL mixed model analysis output.
Program
Monday 12th December
(Optional)
Tuesday 13th December
Wednesday 14th December
Thursday 15th December
Friday 16th December (Optional)
Introduction to Mixed Model Analysis.
Introduction to QTL mapping. Understanding the principles with
the simplest case, a single trait QTL analysis.
 Brief refreshment of quantitative genetic principles.
 Some basics of QTL mapping (marker trait association,
phenotypic variation related to QTLs, conditional QTL
probabilities given marker information).
 Marker regression, simple interval mapping, composite
interval mapping.
Phenotypic analysis of single and multiple trials, the case for
mixed models. Genotype × environment interaction and QTLxE.
 Design and analysis of typical plant breeding trials using
mixed models.
 Genotype by environment analysis. GxE viewed as lack of
parallelism for the mean performance: AMMI, Finlay
Wilkinson regression, factorial regression models. GxE
viewed as heterogeneous variance and lack of correlations
between environments.
 Extension of single environment QTL detection to QTLxE
mixed models.
The use of unstructured genetically diverse population for QTL
detection, the case of association mapping.
 Linkage and linkage disequilibrium.
 The differences/similarities between conventional QTL
mapping and LD mapping.
 The problem of population structure/cryptic relatedness.
 Mixed models for QTL detection accounting for population
structure/cryptic relatedness.
Practices in several software platforms
Costs
Argentine: U$S 100
Others nationalities: U$S 350
Hotels
For help in reservations please contact Lic. Andrea Peña (andreapema@gmail.com).
Information about hotels is attached. Hotel costs are in Argentine Peso, convertion to US
Dollar is 0.236.
Information and registration
Lic. Andrea Peña (andreapema@gmail.com) Cátedra de Estadística y Biometría. Universidad
Nacional de Córdoba. Phone: 54-351-4334105 int 219. Fax 54-351-4334118.
Registration deadline: 5th December, 2011